Kernel panic upon boot

[dmpogo]

It's not NVMe, I'm not rich
But I installed a dist kernel and now it's fine, thanks to you all

Great ! Now you can compare the .config of the dist kernel, and of your attempt for gentoo-sources, and see where the critical difference is.

[pietinger]

Great ! Now you can compare the .config of the dist kernel, and of your attempt for gentoo-sources, and see where the critical difference is.

Hmm ... this is very difficult for a kernel newbie ... because you have to know which are sensible options. In addition, PO wanted to configure a kernel without initramfs. Here, you need to know which modules need to be statically configured into the kernel and for which modules it doesn't matter whether they are static or <M>odules.

I suggest always a "lsmod" (after booting with our Gento dist-kernel) to get an idea of the parts of a machine (of course together with lspci-k). With this list you can search in "make menuconfig" for the description of this module (after hitting / you get the input mask for a search). Select the number and you will be transferred to the option. There you select <Help>.

[dmpogo]

Great ! Now you can compare the .config of the dist kernel, and of your attempt for gentoo-sources, and see where the critical difference is.

Hmm ... this is very difficult for a kernel newbie ... because you have to know which are sensible options. In addition, PO wanted to configure a kernel without initramfs. Here, you need to know which modules need to be statically configured into the kernel and for which modules it doesn't matter whether they are static or <M>odules.

I suggest always a "lsmod" (after booting with our Gento dist-kernel) to get an idea of the parts of a machine (of course together with lspci-k). With this list you can search in "make menuconfig" for the description of this module (after hitting / you get the input mask for a search). Select the number and you will be transferred to the option. There you select <Help>.

Absolutely !

What I tended to do, is to open menuconfig for both kernels side by side, and just go line by line But by now I know the structure of the config, so it is hard to put myself into shoes who does not

[amosbarsinai]

I think that's what I'm gonna do, thank you. I'm not gonna be online though (I'm going back to school on Wednesday... Break's over...)

[Goverp]

A small plug for my cfcfg config comparator, which may, or may not, be of help.

[dmpogo]

A small plug for my cfcfg config comparator, which may, or may not, be of help.

Downloaded

[pietinger]

The only thing I have seen in your output for accessing disks is:

Code: Select all

00:12.7 Universal Flash Storage controller [0109]: Intel Corporation Alder Lake-N UFS Controller [8086:54ff]
        DeviceName: Onboard - Other
        Kernel driver in use: ufshcd

amosbarsinai,

even though you have now installed our Gentoo distribution kernel (and it works), I googled your machine again today while reviewing my threads. And yes, you don't have NVMe, but a highly integrated flash controller. So if you want to try a manual kernel configuration again at some point, you need to enable the following (you may not need all three options, but I don't know your hardware in detail):

Code: Select all

Device Drivers  --->
    [*] Universal Flash Storage Controller  --->
        [*]   Universal Flash Storage BSG device node
        [*]   PCI bus based UFS Controller support
        [*]   Platform bus based UFS Controller support

[amosbarsinai]

For some reason I can only set them as modules... Pressing Enter in nconfig just marks an M

[pietinger]

For some reason I can only set them as modules... Pressing Enter in nconfig just marks an M

I guess you have enabled SCSI as <M>odule ... and it depends on it:

Code: Select all

Depends on: SCSI [=y] && SCSI_DMA [=y]

Set this as built-in ->

Code: Select all

Device Drivers  --->
    SCSI device support  --->
        {*} SCSI device support

[amosbarsinai]

That didn't solve the problem. What you mentioned was already enabled, and I checked again - it isn't able to be *.

[NeddySeagoon]

amosbarsinai,

Something that it depends on will be set as <m>.
Go into the help and read the Depends On:
Everything there must be <*> before you can set <*>

You may need to apply that recursively.

It cannot work as <m>

[amosbarsinai]

Works. Thanks!
Now, should I just run

Code: Select all

make
make modules_install
make install

?
How do I have both the dist-kernel and the manually configured/compiled kernel in GRUB available?

[NeddySeagoon]

amosbarsinai,

Regenerating grub.cfg will do that for you.
You should get both kernels in the menu.
You may need to go into the Advanced menu at the grub menu screen as the kernel you want to boot may not be the default kernel.

[pietinger]

How do I have both the dist-kernel and the manually configured/compiled kernel in GRUB available?

You need your kernel in /boot (either "make install" OR you copy it yourself) and then a grubmkconfig =>
https://wiki.gentoo.org/wiki/User:Pieti ... eat_Sheets
->
Changing the configuration of your used kernel:
->
start after "make menuconfig"

[amosbarsinai]

I entered my linux source directory, ran

Code: Select all

make modules_install
make install
grub-mkconfig -o /boot/grub/grub.cfg

(as root of course)
And for some reason, it installed a new entry called linux-gentoo-gentoo-dist (yeah... it's not a distkernel and there's no reason for the double "gentoo").
So I removed my /boot directory, but then I forgot to rerun grub-install and got a kernel panic...
Long story short, my system is fine now, but I still can't manage to install the manually compiled kernel.

How do I wipe my whole /boot and /efi directories and just install both kernels cleanly?

[pietinger]

[...] How do I wipe my whole /boot and /efi directories and just install both kernels cleanly?

Please dont do this ... remember you have a bootable dist-kernel you can choose in grub ...
(So, you dont need to boot with a bootCD)

[...] but then I forgot to rerun grub-install and got a kernel panic...

1. Dont do a grub install again ... it is not necessary
2. If you have now a kernel panic this means: You have booted your new kernel ... and there is a problem (again) ...

... time to give us your kernel .config to inspect it

... and a "ls -al /boot" ... and to be on a safe side the 10-section of your /boot/grub/grub.cfg (for the moment we dont need more)

[NeddySeagoon]

amosbarsinai,

Your lspci -nnk output would be good too.
If it's further up the topic, that's fine. I didn't look as I'm on my phone just now.

[amosbarsinai]

Uh, sorry, my mistake
Not a kernel panic
A GRUB panic...
The rescue thing?
Something about normal.mod not existing
Anyway I resolved that by booting into the LiveCD, re-emerging the gentoo-kernel-bin, and reinstalling GRUB.
I just want both the distkernel and my own kernel as options...
And what happened up there with the linux-gentoo-gentoo-dist?

[pietinger]

And what happened up there with the linux-gentoo-gentoo-dist?

I dont know ... what do you have on your /boot? (ls -al /boot) ... and the grub.cfg (only 10-section).

Maybe check also these two lines in your .config ->

Code: Select all

#
# General setup
#
[...]
CONFIG_LOCALVERSION=""
# CONFIG_LOCALVERSION_AUTO is not set

What do you have here?

[amosbarsinai]

Code: Select all

╰─>$ grep -i "CONFIG_LOCALVERSION" .config
CONFIG_LOCALVERSION="-gentoo-dist"
# CONFIG_LOCALVERSION_AUTO is not set

Yup, explains that...
But why is CONFIG_LOCALVERSION "-gentoo-dist"?

Code: Select all

╰─>$ ls -al /boot
total 43308
drwxr-xr-x 1 root root      126 Oct 26 20:24 ./
drwxr-xr-x 1 root root      138 Oct 26 20:24 ../
drwxr-xr-x 1 root root       84 Oct 26 20:24 grub/
-rw-r--r-- 1 root root 23808336 Oct 26 20:24 initramfs-6.12.47-gentoo-dist.img
-rw-r--r-- 1 root root 20519920 Oct 26 20:24 kernel-6.12.47-gentoo-dist

I haven't installed my own kernel since I wiped my boot directory...

Code: Select all

╰─>$ sudo head -n 20 /boot/grub/grub.cfg
#
# DO NOT EDIT THIS FILE
#
# It is automatically generated by grub-mkconfig using templates
# from /etc/grub.d and settings from /etc/default/grub
#

### BEGIN /etc/grub.d/00_header ###
if [ -s $prefix/grubenv ]; then
  load_env
fi
if [ "${next_entry}" ] ; then
   set default="${next_entry}"
   set next_entry=
   save_env next_entry
   set boot_once=true
else
   set default="0"
fi

[amosbarsinai]

Never mind, I fixed the problem.
Just needed to NOT wipe my /boot after I saw the weird ending...
Thanks so much to you two, you're always ridiculously helpful!

[pietinger]

Thanks so much to you two, you're always ridiculously helpful!

You are very Welcome!

If I may give you another piece of advice: Install 6.17 right away (it's rock solid now). Then you won't get confused with the versions.

Feel free to write here when your first manually configured kernel boots ... I can still remember how good I felt back then.

I don't know where the “-gentoo-dist” comes from... maybe it's a new default... anyway, get rid of it. (Oh yes, your sound might be difficult ... always these new SoC System-on-a-Chip...) Don't be frustrated and be patient ... yes, you learn that with Gentoo Linux too

If you want to know when I last had a kernel panic ... I'm now on 6.17.5 and my machine has the Intel Volume Management Device [*] ... just take a look here:
https://forums.gentoo.org/viewtopic-t-1 ... ight-.html


But I can reassure you: it happens extremely rarely.

*)

Code: Select all

# lspci -nnk
[...]
0000:00:0e.0 RAID bus controller [0104]: Intel Corporation Volume Management Device NVMe RAID Controller Intel Corporation [8086:a77f]
        DeviceName: Onboard - Other
        Subsystem: Intel Corporation Device [8086:0000]
        Kernel driver in use: vmd

[amosbarsinai]

So...
I emerged sys-kernel/vanilla-sources, enabled everything I was told to in this post, compiled it and installed it, and when I boot it...
Guess what?...
Kernel panic!
Also, I'm not sure if it's a panic... it might just be hanging.
There's no kernel panic message. This happened before (I posted about it here), but even then, the CapsLock button started blinking. Here it all just seems eerily quiet, and stuck on the two first lines.
(The dist and gentoo-sources work fine though, so it's nothing urgent )

[pietinger]

[...]
Kernel panic!
[...]
There's no kernel panic message. [...]

Then it is no kernel panic.

Make sure you have this (statically) enabled to get an output on screen:
https://wiki.gentoo.org/wiki/User:Pieti ... nd_Console

[pietinger]

So...
I emerged sys-kernel/vanilla-sources [...]

So ...
you want to take the hard way.

With gentoo-sources we have enabled some important options automatically with the Kconfig in /usr/src/linux/distro ... I give you an excerpt of this file:

Code: Select all

config GENTOO_LINUX_UDEV
        bool "Linux dynamic and persistent device naming (userspace devfs) support"

        depends on GENTOO_LINUX
        default y if GENTOO_LINUX

        select DEVTMPFS
        select TMPFS
        select UNIX

        select MMU
        select SHMEM

        help
                [...]

config GENTOO_LINUX_PORTAGE
        bool "Select options required by Portage features"

        depends on GENTOO_LINUX
        default y if GENTOO_LINUX

        select CGROUPS
        select NAMESPACES
        select IPC_NS
        select NET_NS
        select PID_NS
        select SYSVIPC
        select USER_NS
        select UTS_NS

config GENTOO_LINUX_INIT_SCRIPT
        bool "OpenRC, runit and other script based systems and managers"

        default y if GENTOO_LINUX

        depends on GENTOO_LINUX

        select BINFMT_SCRIPT
        select CGROUPS
        select EPOLL
        select FILE_LOCKING
        select INOTIFY_USER
        select SIGNALFD
        select TIMERFD